ToxSci Advance Access originally published online on June 12, 2007
Toxicological Sciences 2007 99(1):162-173; doi:10.1093/toxsci/kfm157
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
In Vitro Detection of Drug-Induced Phospholipidosis Using Gene Expression and Fluorescent Phospholipid–Based Methodologies
The Schering-Plough Research Institute, 181 Passaic Avenue, Summit, New Jersey 07901
1 To whom correspondence should be addressed. Fax: (908) 473-7070. E-mail: paul.nioi{at}spcorp.com.
Received May 14, 2007; accepted June 8, 2007
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Phospholipidosis (PLD) is characterized by the excessive intracellular accumulation of phospholipids. It is well established that a large number of cationic amphiphilic drugs have the potential to induce PLD. In the present study, we describe two facile in vitro methods to determine the PLD-inducing potential of a molecule. The first approach is based on a recent study by (Sawada et al., 2005, Toxicol. Sci. 83, 282–292) in which 17 genes were identified as potential biomarkers of PLD in HepG2 cells. To confirm the utility of this gene panel, we treated HepG2 cells with PLD-positive and -negative compounds and then analyzed gene expression using real-time PCR. Our initial analysis, which used a single dose of each drug, correctly identified five of eight positive compounds and four of four negative compounds. We then increased the doses of the three false negatives (amiodarone, tamoxifen, and loratadine) and found that the changes in gene expression became large enough to correctly identify them as PLD-inducing drugs. Our results suggest that a range of concentrations should be used to increase the accuracy of prediction in this assay. Our second approach utilized a fluorescently labeled phospholipid (LipidTox) which was added to the media of growing HepG2 cells along with compounds positive and negative for PLD. Phospholipid accumulation was determined using confocal microscopy and, more quantitatively, using a 96-well plate assay and a fluorescent plate reader. Using an expanded set of compounds, we show that this assay correctly identified 100% of PLD-positive and -negative compounds. Dose-dependent increases in intracellular fluorescent phospholipid accumulation were observed. We found that this assay was less time consuming, more sensitive, and higher throughput than gene expression analysis. To our knowledge, this study represents the first validation of the use of LipidTox in identifying drugs that can induce PLD.
Key Words: phospholipidosis; cationic amphiphilic drugs; gene expression; biomarkers; fluorescent phospholipids; RT-PCR.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Drug-induced phospholipidosis (PLD) is characterized by the excessive intracellular accumulation of phospholipids and by the appearance of membranous lamellar inclusions known as lamellar bodies. There are over 50 marketed pharmaceuticals that induce PLD and these are all cationic amphiphilic drugs (CADs) that are characterized by a hydrophobic ring structure and a hydrophilic side chain with a charged amine group. The mechanism of drug-induced PLD is thought to involve direct binding of CADs to phospholipids creating a complex that is resistant to degradation by phospholipases (Reasor et al., 2006
). There is also evidence that suggests some CADs can cause PLD by directly inhibiting phospholipase activity (Halliwell, 1997; Reasor et al., 2006).
Although a large number of approved drugs can induce PLD, there is no strong evidence that this condition is harmful to human health. It is also important to realize that PLD is reversible, so once drug treatment is terminated, the CAD is eventually expelled from the cell and phospholipid levels return to normal. The kinetics of this reversal is dependent upon the CAD in question, but in preclinical species this process spans several weeks (Reasor et al., 2006).
From a regulatory perspective, drug-induced PLD is considered an adverse event for three main reasons. Firstly, there are several genetic conditions associated with PLD, the most frequently cited of which is Niemann-Pick disease which results in the abnormal accumulation of sphingomyelin in various cell types (Mcgovern et al., 2006). This disease can prove fatal depending upon the exact mutations present in a given case. Although the biology of Niemann-Pick disease is not fully understood, there is clearly a link between abnormal intracellular phospholipid accumulation and mortality. Secondly, there is the issue of understanding the relationship of PLD to other toxicity observed in the same tissue. For example, gentamycin causes PLD in kidney tissue and this is associated with renal tubular toxicity (Kaloyanides and Pastoriza-Munoz, 1980; Laurent et al., 1990). It is not clear whether the PLD is the cause of the toxicity or if it is just another cellular change seen in the same tissue that is unrelated to the tubular damage. Finally, despite the fact that PLD has not been shown to be harmful to humans, there is still uncertainty over the consequences of drug-induced PLD on cell and tissue function. It is common for the Food and Drug Administration (FDA) to request additional studies to be conducted on tissues where PLD is seen in order to show that cellular function remains unaltered. In 2004, the FDA established a PLD working group whose objective is to combine data from a variety of sources to determine whether compounds that induce PLD pose a clinical risk.
It is important to obtain as much information about the potential liabilities of compounds in development as early as possible. Therefore, it would be advantageous to have a rapid in vitro screen to check if any given compound could induce PLD. This information could then be used as part of investigative toxicology efforts to proactively establish whether the PLD had any effect on cell or tissue function. The gold standard method of detecting PLD is electron microscopy, which can be used to identify lamellar bodies in CAD-treated cells. However, this methodology is not conducive to screening large numbers of compounds. A variety of other in vitro methodologies have recently been described which use fluorescent dyes (e.g., Nile red, Casertelli et al., 2003) or fluorescently labeled phospholipids (e.g., 1-acyl-2-[12-(7-nitro-2,1,3-benzoxadiazol-4-yl)amino]dodecanoyl]-sn-glycerol-3-phosphocholine [NBD-PC] Kasahara et al., 2006) in cell culture to assess PLD. In addition, Sawada et al. (2005) recently identified a panel of 17 genes, the expressions of which either was up- or downregulated in response to PLD-inducing drugs. To increase the throughput of this assay, it was transferred to a 96-well plate high-throughput genomics–based platform (Sawada et al., 2006).
The present study had two aims (1) to assess if we could use the 17 gene panel described by Sawada et al. (2005) to identify drugs that induce PLD and (2) to validate the utility of a novel fluorescently labeled phospholipid (LipidTox), which has not been described in the literature, in determining if a compound causes PLD.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Chemicals.
Unless otherwise stated, all chemicals were purchased from either Sigma or MP Biomedicals. Information on each of the PLD-inducing drugs used can be found in Table 1. LipidTox reagent was purchased from Invitrogen.
|
Cell culture.
HepG2 cells were maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum, both of which were obtained from Invitrogen.
RNA isolation and quantitative PCR.
For real-time PCR analysis, HepG2 cells were seeded in 12-well plates at a density of 4x105 cells per well, allowed to adhere overnight, and then treated for a total of 24 h with each compound. RNA was then isolated using an RNeasy mini kit (Qiagen) according to the manufacturer's instructions. The quality and concentration of RNA was determined using an Agilent 2100 bioanalyzer (Agilent)—all RNA used in this study achieved a ribosomal integrity number of 9.0 or greater. A total of 2 µg of RNA was reversed transcribed using the TaqMan reverse transcription kit (Applied Biosystems) according to the manufacturer's instructions. Quantitative PCR was conducted using a reaction containing 20 ng of cDNA (based on the assumption that conversion of RNA to cDNA was 100% successful) 1x SYBR green PCR master mix (Applied Biosystems) and either 100nM (18S RNA) or 200nM (all other genes) forward and reverse oligonucleotides. All oligonucleotide sequences have been previously published (Sawada et al., 2005) apart from those used for ASAH1. In this case, the forward oligo was 5'-ACCCTAAGGAAGTTGCTAACTTAAAAAA-3', and the reverse oligo was 5'-ACTAAATTAACAGAACGTGGGATGC-3'. Samples were assembled using a Biomek NX robot (Beckman Coulter) and analyzed using an Applied Biosystems 7900HT real-time PCR system (Applied Biosystems). 18S RNA concentration was used to normalize the amount of RNA present in each sample.
Confocal microscopy.
For confocal analysis, cells were seeded into six-well plates, containing coverslips, at a density of 1 x 106 cells per well. Cells were allowed to attach overnight and were then treated with LipidTox (according to the manufacturer's instructions) and each of the compounds used in the study. After 48 h, cells were fixed in 3% (vol/vol) formaldehyde in phosphate-buffered saline (PBS) containing 10 µg/ml Hoechst 33258 for 30 min. Following washing, coverslips were mounted on glass slides using Prolong Gold antifade reagent (Invitrogen). Confocal images were obtained using a Leica SP5 microscope; LipidTox was excited at 543 nm and its emission was detected at 594 nm, while Hoechst was excited at 405 nm and its emission was detected at 490 nm.
Microplate assay.
HepG2 cells were seeded into 96-well plates at a density of 3 x 104 cells per well and allowed to attach overnight. Cells were then treated in triplicate with LipidTox (according to the manufacturer's instructions) and each of the compounds tested in the study. After 48 h, cells were fixed in 3% (vol/vol) formaldehyde in PBS containing 10 µg/ml Hoechst 33258 for 30 min. Following washing, fluorescence was measured using a Spectrafluor Plus microplate reader (Tecan). LipidTox fluorescence was detected using a 530 nm excitation filter and a 635 nm emission filter, whereas Hoechst fluorescence was determined using a 360 nm excitation filter and a 465 nm emission filter. Values for LipidTox fluorescence were normalized to those of Hoechst fluorescence to control for cell number. Statistical analyses were performed using GraphPad Prism software (GraphPad software, Inc.).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Treatment of HepG2 Cells with Some, but not all, Compounds that Induce PLD Results in the Differential Expression of a Subset of Genes
Recently, Sawada et al. (2005
) identified a panel of 17 genes that are induced or repressed when HepG2 cell cultures are exposed to drugs that induce PLD (Table 2). In order to confirm the utility of this gene panel as a predictor of PLD-inducing potential, HepG2 cells were treated with eight known inducers of PLD (amiodarone, amitriptyline, fluoxetine, imipramine, ketoconazole, sertraline, tamoxifen, and loratadine) at doses that have been shown to cause the formation of lamellar bodies in these cells (Table 1). We also treated with two CADs (citalopram and doxepin) which, based on their structures and information available in the literature, are likely to cause PLD but have not been shown to do so in HepG2 cells. Acetaminophen, erythromycin, sotalol, and quinidine were used as negative controls at concentrations previously shown not to cause PLD in HepG2 cells (Table 1). It is important to note that both erythromycin and quinidine are known to cause PLD in vivo and in vitro in other cell types (Table 1). Cells were exposed to drug for 24 h, then RNA was isolated from them, reverse transcribed, and analyzed by quantitative PCR using primers specific for the genes listed in Table 2. The results of this analysis are shown in Table 3. Generally, the negative control compounds had little or no effect on the expression of any of the genes examined, although quinidine did cause induction/repression of a few (this may be related to the fact that quinidine can cause PLD in other cell types). By contrast, exposure of cells to known inducers of PLD caused substantial changes in the expression of the genes in the panel. Some genes responded better than others for example, FABP1, p8, SERPINA3, and TAGLN were all relatively strongly induced or repressed, whereas NROB2 and PHYH showed no response (Table 3). For any given gene, the level of increase or decrease in its corresponding mRNA was different depending upon the compound used, for example, amitriptyline had relatively strong effects, whereas amiodarone and loratadine were much weaker. We analyzed these data further by calculating the phospholipidotic index (PI) score for each compound, which is the average fold change in gene expression for the gene panel (Table 3). We defined a positive result (i.e., that which indicates a compound can cause PLD) as a PI score 
1.5, which is consistent with the original study by Sawada et al. (2005). Furthermore, in our hands, a fold change in gene expression = 1.5 represents the threshold where we have confidence that this value is significantly different from 1.0.
|
|
All the negative control compounds had a PI score < 1.5 and so were correctly identified (Table 3). For the positive drugs, five of eight had a PI score
1.5, but amiodarone, loratadine, and tamoxifen all had scores < 1.5 (Table 3). Citalopram and doxepin were both identified as positive in our analysis, which is consistent with what we expected based upon the literature and the structures of these drugs (Table 3).
When we repeated this experiment, we obtained PI scores that were very consistent with our initial study (Table 4). However, this did reveal that compounds with PI scores close to 1.5 are called incorrectly more often than those with higher scores. For example, in experiment 1, citalopram and doxepin gave values of 1.66 and 1.76, respectively, but in experiment 2 they gave PI scores of 1.43 and 1.33, respectively (Table 4). We also compared our PI scores with those from the study by Sawada et al. (2005) and from another study which was published while this manuscript was in preparation (Atienzar et al., 2006). Our results are consistent with the previously published data with a couple of notable differences (Table 4). Sawada et al. (2005) identified amiodarone, loratadine, and tamoxifen as PLD-inducing drugs with PI scores of 1.61, 1.59, and 2.15, respectively, whereas in our study all of these compounds had PI scores < 1.5 (Table 4). Consistent with our data, Atienzar et al. (2006) also generated PI scores of < 1.5 for amiodarone and tamoxifen; they did not examine loratadine (Table 4). The only other major difference in the three data sets was with fluoxetine which had a PI score consistent with it inducing PLD in our study (1.99) and in Sawada's study (3.29) but not in Atienzar's experiment (1.27) (Table 4).
|
Treatment of Cells with Increasing Doses of PLD-Inducing Drugs Results in a Concomitant Increase in the PI Score
As mentioned above, amiodarone, loratadine, and tamoxifen were not correctly identified as PLD-inducing drugs. We wanted to determine if the level of change in gene expression would become large enough to give a PI score
1.5 if cells were exposed to higher concentrations of these compounds. HepG2 cells were treated for 24 h with 8.3, 16, and 25µM amiodarone and loratadine, 8.3 and 16µM tamoxifen (higher concentrations are toxic), and 25, 50 and 75µM acetaminophen as a negative control. As in the previous experiment, total RNA was isolated from treated cells, and the levels of transcripts corresponding to the 17 genes in the panel were determined by real-time PCR (Table 5). We calculated the PI scores for each of the treatments and found that they increased as the doses of amiodarone, loratadine, and tamoxifen were increased but were unaffected by elevating the concentration of acetaminophen (Table 5). PI scores of 1.5 were achieved for amiodarone at 16µM (1.53) and 25µM (2.35), for loratadine at 25µM (1.61), and for tamoxifen at 16µM (1.81). In contrast, even at 75µM acetaminophen had a PI score of 1.23 (Table 5).
|
Treatment of HepG2 Cells with PLD-Inducing Drugs Causes the Intracellular Accumulation of a Fluorescently Labeled Phospholipid
As an alternative to the above approach, we utilized a novel fluorescently labeled phospholipid (LipidTox) to monitor drug-induced PLD in HepG2 cells. The formulation of LipidTox is proprietary, but it is distinct from other fluorescently labeled phospholipids (such as NBD-PE) that are commercially available. Given that distinct types of phospholipid may accumulate differently in PLD (Diez-Blanco et al., 1987
), it is important to validate LipidTox especially since to our knowledge there has been no published data using this reagent.
HepG2 cells grown on glass coverslips were cotreated with LipidTox and each of the compounds used in our initial real-time PCR study (Tables 1 and 3). The concentrations used were also the same as those used in the experiment shown in Table 3. After 48 h, cells were fixed, stained with Hoechst, and imaged by confocal microscopy. LipidTox and Hoechst emit red and blue fluorescence, respectively. Representative images are shown in Figure 1A. When HepG2 cells were treated with seven of eight known inducers of PLD (amiodarone, amitriptyline, fluoxetine, imipramine, ketoconazole, sertraline, and tamoxifen), a substantial increase in red fluorescence was observed, indicating phospholipid accumulation (Fig. 1A). Loratadine was the only positive control compound that did not cause a noticeable increase in red fluorescence. Citalopram and doxepin also caused marked intracellular phospholipid accumulation (Fig. 1A) which is consistent with data from other cell lines and with their structures (Table 1). In contrast, cells treated with either acetaminophen or sotalol showed no increase in red fluorescence over vehicle-treated cells which is consistent with the literature (Fig. 1 and Table 1). Interestingly, when cells were treated with either erythromycin or quinidine, a noticeable increase in intracellular phospholipid was observed (Fig. 1A). Both of these compounds are well-established inducers of PLD in other cell lines and in vivo (Table 1), although Sawada et al. (2005) did not observe the formation of lamellar bodies in HepG2 cells treated with these drugs. This could be due to variation in the culture conditions used or in the HepG2 cells themselves, but our data are more consistent with the wider literature (Table 1).
|
In order to obtain more quantitative data and develop a higher throughput system, we adapted this assay to a 96-well plate format. HepG2 cells were treated and fixed as described above, and fluorescence was measured using a microplate reader. The fluorescent intensity of LipidTox was normalized to that of Hoechst to control for cell number. We did not observe any large differences in Hoechst fluorescence at the concentrations of compounds that we tested. Good agreement between the confocal images and microplate readings was seen (Figs. 1A and 1B). Treatment of cells with amiodarone, amitriptyline, fluoxetine, imipramine, ketoconazole, sertraline, tamoxifen, citalopram, doxepin, or quinidine gave a statistically significant increase in relative fluorescence (Fig. 1B). The results obtained from erythromycin-treated cells were also elevated, although this was not statistically significant. Exposure of HepG2 cells to loratadine did not cause an increase in relative fluorescence over Dimethyl sulfoxide–treated cells (Figs. 1A and 1B, Table 1), which is consistent with our confocal data but at odds with the fact that this compound is a known inducer of PLD. As expected, the negative control compounds acetaminophen and sotalol did not cause increased accumulation of LipidTox in the cells (Fig. 1B).
Treatment of HepG2 Cells with Increasing Concentrations of PLD-Inducing Drugs Causes a Parallel Increase in the Accumulation of Fluorescent Phospholipid
In order to determine if increasing the dose of drugs that induce PLD would also cause an increase in the intracellular accumulation of LipidTox, HepG2 cells were treated for 48 h with LipidTox and with 8.3, 16, and 25µM loratadine and ketoconazole, 25, 50, 100, and 200µM erythromycin, and 25, 50, and 100µM acetaminophen. Cells were fixed, stained with Hoechst, and analyzed by confocal microscopy (Fig. 2). As the concentrations of loratadine, ketoconazole, and erythromycin were increased, the level of intracellular red fluorescence also got higher (Fig. 2). However, even when cells were treated with 100µM acetaminophen, no accumulation of the phospholipid could be seen (Fig. 2).
|
To quantitate and expand this data, we used the 96-well plate–based assay. HepG2 cells were cotreated for 48 h with LipidTox and with increasing doses of all the drugs used in this study. After fixation, fluorescence was read on a microplate reader, and the results are shown in Figure 3. A dose-dependent increase in red fluorescence was seen for all of the compounds except the negative controls acetaminophen and sotalol. At the maximum dose of each PLD-inducing drug used, the increase in fluorescence was statistically significant. The results for loratadine and erythromycin are particularly noteworthy since both of the compounds are known inducers of PLD, but when used at their lowest concentrations in this study, little or no phospholipid accumulation could be seen (Fig. 3). However, as stated above, as the concentrations of these drugs were increased, a statistically significant increase in intracellular LipidTox was detected (Fig. 3).
|
Differences were observed between the compounds, both in terms of the minimum dose required to give a positive result and in the maximum response attainable. For example, at a concentration of 8.3µM loratadine, citalopram and doxepin gave results similar to control-treated cells, whereas at the same dose, a number of compounds including amitriptyline, fluoxetine, and sertraline were clearly positive for PLD (Fig. 3). The maximum effect of amiodarone was seen at a concentration of 4µM (around a twofold increase in LipidTox accumulation), and this effect could not be enhanced by further increasing the dose of the drug (Fig. 3). By contrast, adding additional amounts of amitriptyline, fluoxetine, and sertraline caused more accumulation of LipidTox in the cell (Fig. 3).
Further Validation of the LipidTox Assay Using an Additional Set of Compounds
To validate further the fluorescent phospholipid–based assay, we used an additional set of 10 compounds; chlorpromazine, perhexiline, clozapine, clomipramine, and thioridazine are all known inducers of PLD, whereas ofloxacin, procainamide, phenobarbital, valproic acid, and tert-butylhydroquinone are not (Table 1). HepG2 cells were treated with LipidTox and exposed to a single dose of the above drugs for 48 h. Cells were analyzed by confocal microscopy as described above. All five of the positive control compounds caused a significant increase in intracellular red fluorescence indicating PLD, but no increase was observed with the negative control drugs (Fig. 4). Again, we used the microplate assay to further define these responses. Cells were treated for 48 h with LipidTox and increasing doses of all of the compounds and then processed as described above. A dose-dependent increase in relative fluorescence was observed in cells treated with the positive control drugs (Fig. 5). However, even at high doses, none of the negative controls caused the accumulation of LipidTox in the HepG2 cells (Fig. 5).
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Over 50 approved drugs can cause PLD, but there is no evidence that this condition is harmful to humans. However, for a number of reasons, the regulatory authorities consider PLD to be an adverse event, and this can considerably delay or prevent the further development of a compound. The present study aimed to develop an in vitro screen that could be used to identify drugs that cause PLD. This would allow proactive investigative studies to be carried out to determine if the PLD had any effect on cellular function.
Our first strategy was to determine if we could confirm the use of the set of 17 genes identified by Sawada et al. (2005) to detect the potential of a drug to cause PLD. We found that only five of eight of the positive control compounds were correctly identified when a single concentration of drug was used, and these findings were consistent between experiments. We used the same cell line and the same concentration of drug as Sawada et al. (2005), but they were able to identify the false negatives from our experiment (amiodarone, loratadine, and tamoxifen) as positive. During the preparation of this manuscript, Atienzar et al. (2006) published data consistent with ours, showing that both amiodarone and tamoxifen (they did not examine loratadine) were false negatives. However, when we increased the concentrations of amiodarone, loratadine, and tamoxifen, the changes in gene expression became large enough to correctly identify them as positive. We suggest that a range of concentrations of each drug tested should be used to reduce the number of false negative results. All the negative control compounds we tested (four of four) gave PI scores < 1.5. Two of these drugs, erythromycin and quinidine, are known to cause PLD in vivo and in other cell lines (Table 1), but Sawada et al. (2005) could not identify lamellar bodies in HepG2 cells treated with these compounds at the doses used in our study. It would be interesting to see if increasing the concentration of these drugs would give a PI score 1.5.
Not all of the genes in the panel responded well to CAD treatment; the expression of NROB2 and PHYH was not induced by any of the positive control drugs used in the study. It might be possible to remove these genes from the panel, although there is still a possibility that certain compounds that were not tested in our study could provoke a strong transcriptional response from these genes. These genes were also weak responders in the data published by Atienzar et al. (2006), but Sawada et al. (2005) found that the expression of NROB2 correlated well with PLD while that of PHYH did not. This is an interesting difference and might indicate that the transcriptional pathways that are available for activation in the HepG2 cells used by Sawada et al. (2005) are different from those present in the HepG2 cells that we used. This hypothesis is supported by the fact that a study has shown that there is heterogeneity between HepG2 sublines used in different laboratories (Van Pelt et al., 2003).
The role of the genes induced or repressed in PLD remains unclear. Current dogma suggests that the development of PLD is related to the inhibition of phospholipase activity either through the generation of CAD-phospholipid complexes or by direct inhibition of phospholipase activity. If this is the case, the genes we examined probably encode proteins that are induced as a consequence of the presence of PLD in the cell. A number of the genes encode proteins that have a role in cholesterol metabolism, the handling of phospholipids, or in maintaining energy balance. For example, ASAH1 is a protein that functions in lysosomes and catalyzes the breakdown of ceramide into sphingosine and fatty acids (Park and Schuchman, 2006). Ceramide can be synthesized from the phospholipid sphingomyelin by sphingomyelin phosphodiesterase. It is possible that drugs that cause PLD result in the intracellular accumulation of sphingomyelin which may in turn lead to increased levels of ceramide. Thus, induction of ASAH1 may represent an adaptive response to reduce intracellular levels of ceramide. In contrast to ASAH1, which was induced by drugs that cause PLD, SLC2A3 was repressed. SLC2A3 is also known as GLUT3 and functions as a glucose transporter. Again, the precise reason for the altered expression of this gene is not known; however, decreased expression of another glucose transporter (GLUT4) has been shown to increase lipid production in the liver (Kotani et al., 2004). If enhanced production of lipid can decrease glucose transporter expression, then it would be interesting to examine whether the accumulation of phospholipids could have a similar effect. A lack of available glucose is known to drive cells to increase lipid production for use as fuel. Therefore, repression of SLC2A3 could represent an adaptive response to high intracellular lipid levels that could be used preferentially to fulfill energy requirements. Other genes encode proteins that have functions that cannot be easily linked to PLD (Table 2). For example, p8 is a protein that has opposing functions; on one hand, it can cause apoptosis and on the other, it was shown to promote cell proliferation and is essential for the growth of certain tumors (Carracedo et al., 2006; Vasseur et al., 2002). A recent study demonstrated that p8 was induced by ceramide and subsequently caused the cells in the study to enter apoptosis (Carracedo et al., 2006). Interestingly, several PLD-inducing drugs have been shown to cause apoptosis in cell lines (Piccotti et al., 2005). As described above, it is possible that ceramide levels are increased in phospholipidotic cells, which could explain the induction of p8. Clearly, careful analysis of the exact function of p8 in PLD is needed as this may or may not indicate a potential adverse effect of phospholipid accumulation in cells. This highlights the concerns of conducting gene expression analysis since data can be generated that is not easily interpretable.
Because of the concerns highlighted above, we wanted to develop a more direct and easily interpretable method for assessing PLD. We used LipidTox, which is a fluorescently labeled phospholipids, to assess the induction of PLD in HepG2 cells. Using this system, we were able to identify compounds that induce PLD in vivo with 100% accuracy. Interestingly, erythromycin and quinidine are both known to cause PLD in vivo and in vitro but were not shown to do so in HepG2 cells at concentrations of 25µM (Table 1). These compounds were not identified as positive in our gene expression analysis, but even at a concentration of 25µM in the LipidTox assay, intracellular phospholipid accumulation could be seen. The degree of PLD was increased in a dose-dependent manner by these drugs. The results indicate that the LipidTox assay is more sensitive than gene expression analysis for detecting PLD in HepG2 cells. We also found that amiodarone and tamoxifen were easily identified as causing PLD at concentrations that gave PI scores < 1.5 in the gene expression study. This is significant because as the dose of these drugs is increased, more toxicity can be seen. For example, tamoxifen was used at a maximum dose of 16µM because most of the cells were dead at concentrations higher than this. We were also able to adapt the LipidTox assay to a 96-well plate–based format to allow for higher throughput screening of compounds.
The LipidTox assay has two main advantages over gene expression analysis. Firstly, it is more sensitive and can more accurately identify compounds capable of causing PLD at lower concentrations. Secondly, it is more rapid and higher throughput. As with the testing of any scientific hypothesis, consistent data from multiple approaches provides the most convincing results. It is conceivable that the LipidTox assay could be used to identify PLD-inducing drugs from a large series of compounds, and then confocal microscopy and gene expression analysis could be used to add weight to the results. The main caveats of these approaches are related to the metabolic profile of HepG2 cells. If a parent compound is capable of inducing PLD but is rapidly transformed to an inactive metabolite in HepG2 cells, then it will not be correctly identified. The converse is also true; i.e., if a metabolite of the parent compound can cause PLD but HepG2 cells lack the enzymes needed for this transformation, then it will not be picked up as positive.
To our knowledge, this study provides the first validation of the use of LipidTox reagent to detect PLD. This is important because it is known that there are differences in the accumulation of distinct phospholipids in PLD (Diez-Blanco et al., 1987). Another recent study established a 96-well plate–based screening assay for detecting compounds that induce PLD, but they used NBD-PC as their fluorescent phospholipid (Kasahara et al., 2006). There are a number of significant differences between the results of this study and ours. Erythromycin, loratadine, and quinidine were not correctly identified as PLD-inducing drugs even at concentrations of 30µM in their assay (Kasahara et al., 2006). In contrast, all three drugs were identified as positive in our experiment at concentrations of 25µM. These differences could be a result of the use of different cell lines in each of the studies (CHO-K1 vs. HepG2) or might be related to the different phospholipids used in each experiment. A comparison of LipidTox accumulation in a range of cell lines treated with the same compounds would help to address these issues.
In conclusion, the present study confirmed the utility of a gene expression approach in determining the PLD-inducing potential of a drug. Our results suggest that a number of concentrations of a drug should be used to increase the accuracy of prediction. We have also developed a fluorescent phospholipid-based assay for detecting PLD. This system allowed us to identify compounds capable of inducing PLD with 100% accuracy. We were able to adapt this assay to a 96-well plate-based format to allow higher throughput screening for drugs that cause PLD. The LipidTox assay offers several advantages over gene expression analysis, and we would suggest that this should be the method of choice for screening for compounds that cause PLD.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Atienzar F, Gerets H, Dufrane S, Tilmant K, Cornet M, Dhalluin S, Ruty B, Rose G, Canning M. Determination of phospholipidosis potential based on gene expression analysis in HepG2 cells. Toxicol. Sci. (2006) 96:101–114.[CrossRef][Web of Science][Medline]
Carracedo A, Lorente M, Egia A, Blazquez C, Garcia S, Giroux V, Malicet C, Villuendas R, Gironella M, Gonzalez-Feria L, et al. The stress-regulated protein p8 mediates cannabinoid-induced apoptosis of tumor cells. Cancer Cell (2006) 9:301–312.[CrossRef][Web of Science][Medline]
Casartelli A, Bonato M, Cristofori P, Crivellente F, Dal Negro G, Masotto I, Mutinelli C, Valko K, Bonfante V. A cell-based approach for the early assessment of the phospholipidogenic potential in pharmaceutical research and drug development. Cell Biol. Toxicol. (2003) 19:161–176.[CrossRef][Web of Science][Medline]
Diez-Blanco MN, Sanchez-Yague J, Cabezas JA, Llanillo M. Isolation, characterization and phospholipid composition of lamellar bodies and subcellular fractions from dog lung. Int. J. Biochem. (1987) 19:693–698.[CrossRef][Web of Science][Medline]
Drenckhahn D, Kleine L, Lullmann-Rauch R. Lysosomal alterations in cultured macrophages exposed to anorexigenic and psychotropic drugs. Lab. Investig. (1976) 35:116–123.[Web of Science][Medline]
Drew R, Siddik ZH, Mimnaugh EG, Gram TE. Species and dose differences in the accumulation of imipramine by mammalian lungs. Drug Metab. Dispos. (1981) 9:322–326.[Abstract]
Gonzalez-Rothi RJ, Zander DS, Ros PR. Fluoxetine hydrochloride (Prozac)-induced pulmonary disease. Chest (1995) 107:1763–1765.[Web of Science][Medline]
Halliwell WH. Cationic amphiphilic drug-induced phospholipidosis. Toxicol. Pathol. (1997) 25:53–60.
Hansson AL, Xia Z, Berglund MC, Bergstrand A, Depierre JW, Nässberger L. Reduced cell survival and morphological alterations induced by three tricyclic antidepressants in human peripheral monocytes and lymphocytes and in cell lines derived from these cell types. Toxicol. In Vitro (1997) 11:21–31.[CrossRef][Web of Science]
Honegger UE, Zuehlke RD, Scuntaro I, Schaefer MH, Toplak H, Wiesmann UN. Cellular accumulation of amiodarone and desethylamiodarone in cultured human cells. Consequences of drug accumulation on cellular lipid metabolism and plasma membrane properties of chronically exposed cells. Biochem. Pharmacol. (1993) 45:349–356.[CrossRef][Web of Science][Medline]
Jaeschke H, Gores GJ, Cederbaum AI, Hinson JA, Pessayre D, Lemaster JJ. Mechanisms of hepatotoxicity. Toxicol. Sci. (2002) 65:166–176.
Jagel M, Lullmann-Rauch R. Lipidosis-like alterations in cultured macrophages exposed to local anaesthetics. Arch. Toxicol. (1984) 55:229–232.[CrossRef][Web of Science][Medline]
Kaloyanides GJ, Pastoriza-Munoz E. Aminoglycoside nephrotoxicity. Kidney Int. (1980) 18:571–582.[Web of Science][Medline]
Kasahara T, Tomita K, Murano H, Harada T, Tsubakimoto K, Ogihara T, Ohnishi S, Kakinuma C. Establishment of an in vitro high-throughput screening assay for detecting phospholipidosis-inducing potential. Toxicol. Sci. (2006) 90:133–141.
Kodavanti UP, Lockard VG, Mehendale HM. In vivo toxicity and pulmonary effects of promazine and chlorpromazine in rats. J. Biochem. Toxicol. (1990) 5:245–251.[CrossRef][Web of Science][Medline]
Kotani K, Peroni OD, Minokoshi Y, Boss O, Kahn BB. GLUT4 glucose transporter deficiency increases hepatic lipid production and peripheral lipid utilization. Clin. Investig. (2004) 114:1666–1675.[CrossRef]
Laurent G, Kishore BK, Tulkens PM. Aminoglycoside-induced renal phospholipidosis and nephrotoxicity. Biochem. Pharmacol. (1990) 40:2383–2392.[CrossRef][Web of Science][Medline]
Lullmann-Rauch R. Lipidosis-like ultrastructural alterations in rat lymph nodes after treatment with tricyclic antidepressants or neuroleptics. Naunyn Schmiedebergs Arch. Pharmacol. (1974) 286:165–179.[CrossRef][Web of Science][Medline]
Lullmann-Rauch R, Nassberger L. Citalopram-induced generalized lipidosis in rats. Acta Pharmacol. Toxicol. (copenh.) (1983) 52:161–167.[Medline]
McGovern MM, Aron A, Brodie SE, Desnick RJ, Wasserstein MP. Natural history of Type A Niemann-Pick disease: Possible endpoints for therapeutic trials. Neurology (2006) 66:228–232.
Montenez JP, Van Bambeke F, Piret J, Brasseur R, Tulkens PM, Mingeot-Leclercq MP. Interactions of macrolide antibiotics (Erythromycin A, roxithromycin, erythromycylamine [Dirithromycin], and azithromycin) with phospholipids: Computer-aided conformational analysis and studies on acellular and cell culture models. Toxicol. Appl. Pharmacol. (1999) 156:129–140.[CrossRef][Web of Science][Medline]
Morelli JK, Buehrle M, Pognan F, Barone LR, Fieles W, Ciaccio PJ. Validation of an in vitro screen for phospholipidosis using a high-content biology platform. Cell Biol. Toxicol. (2006) 22:15–27.[CrossRef][Web of Science][Medline]
Pakuts AP, Parks RJ, Paul CJ, Bujaki SJ, Mueller RW. Ketoconazole-induced hepatic lysosomal phospholipidosis: The effect of concurrent barbiturate treatment. Res. Commun. Chem. Pathol. Pharmacol. (1990) 67:55–62.[Web of Science][Medline]
Park JH, Schuchman EH. Acid ceramidase and human disease. Biochim. Biophys. Acta (2006) 1758:2133–2138.[Medline]
Pessayre D, Bichara M, Degott C, Potet F, Benhamou JP, Feldmann G. Perhexiline maleate-induced cirrhosis. Gastroenterology (1979) 76:170–177.[Web of Science][Medline]
Piccotti JR, LaGattuta MS, Knight SA, Gonzales AJ, Bleavins MR. Induction of apoptosis by cationic amphiphilic drugs amiodarone and imipramine. Drug Chem. Toxicol. (2005) 28:117–133.[CrossRef][Web of Science][Medline]
Reasor MJ. A review of the biology and toxicologic implications of the induction of lysosomal lamellar bodies by drugs. Toxicol. Appl. Pharmacol. (1989) 97:47–56.[CrossRef][Web of Science][Medline]
Reasor MJ, Davis ME. Prevention of chlorphentermine-induced pulmonary phospholipidosis in rats by phenobarbital. Drug Metab. Dispos. (1985) 13:192–196.[Abstract]
Reasor MJ, Hastings KL, Ulrich RG. Drug-induced phospholipidosis: Issues and future directions. Expert Opin. Drug Saf. (2006) 5:567–583.[CrossRef][Web of Science][Medline]
Reasor MJ, Kacew S. Drug-induced phospholipidosis: Are there functional consequences? Exp. Biol. Med. (2001) 226:825–830.
Sawada H, Takami K, Asahi S. A toxicogenomic approach to drug-induced phospholipidosis: Analysis of its induction mechanism and establishment of a novel in vitro screening system. Toxicol. Sci. (2005) 83:282–292.
Sawada H, Taniguchi K, Takami K. Improved toxicogenomic screening for drug-induced phospholipidosis using a multiplexed quantitative gene expression ArrayPlate assay. Toxicol. In Vitro (2006) 20:1506–1513.[Web of Science][Medline]
Tang W, Borel AG, Fujimiya T, Abbott FS. Fluorinated analogues as mechanistic probes in valproic acid hepatotoxicity: Hepatic microvesicular steatosis and glutathione status. Chem. Res. Toxicol. (1995) 8:671–682.[CrossRef][Web of Science][Medline]
Van Pelt JF, Decorte R, Yap PS, Fevery J. Identification of HepG2 variant cell lines by short tandem repeat (STR) analysis. Mol. Cell. Biochem. (2003) 243:49–54.[CrossRef][Web of Science][Medline]
Vasseur S, Hoffmeister A, Garcia S, Bagnis C, Dagorn JC, Iovanna JL. p8 is critical for tumour development induced by rasV12 mutated protein and E1A oncogene. EMBO Rep. (2002) 3:165–170.[CrossRef][Web of Science][Medline]
Waters E, Wang JH, Redmond HP, Wu QD, Kay E, Bouchier-Hayes D. Role of taurine in preventing acetaminophen-induced hepatic injury in the rat. Am. J. Physiol. Gastrointest. Liver Physiol. (2001) 280:G1274–G1279.
Whitehouse LW, Menzies A, Mueller R, Pontefract R. Ketoconazole-induced hepatic phospholipidosis in the mouse and its association with de-N-acetyl ketoconazole. Toxicology (1994) 94:81–95.[CrossRef][Web of Science][Medline]
Xia Z, Ying G, Hansson AL, Karlsson H, Xie Y, Bergstrand A, DePierre JW, Nassberger L. Antidepressant-induced lipidosis with special reference to tricyclic compounds. Prog. Neurobiol. (2000) 60:501–512.[CrossRef][Web of Science][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||